Evaluation of Classification Algorithms for Phishing URL Detection

Volume 6 Issue 3 September - November 2018

Research Paper

Evaluation of Classification Algorithms for Phishing URL Detection

Oluyomi Ayanfeoluwa *, Oluwafemi Osho**, Maryam Shuaib***

* President, Information System Audit & Control Association (ISACA), Federal University of Technology, Minna, Nigeria.

** Lecturer, Department of Cyber Security Science, Federal University of Technology Minna, Nigeria.

*** Former Special Assistant, ICT Development to the Governer of Nigeria State, Nigeria.

Oluyomi, A., Osho, O., Shuaib, M.(2018) Evaluation of Classification Algorithms for Phishing URL Detection, i-manager's Journal on Computer Science 6(3),34-41. https://doi.org/10.26634/jcom.6.3.15698

Abstract

A phishing URL is a web address created with the intent of deceiving users into releasing their personal and private data or downloading malware into the users' systems without their knowledge. Increase in the adoption of the Internet has led to corresponding increase in the number of phishing sites globally. Many classification techniques have been developed for detecting phishing URLs. This paper seeks to evaluate the performances of existing techniques. With dataset obtained from UCI Machine Learning Repository, the algorithms were assessed in terms of Accuracy, Precision, Recall, F-Measure, Receiver Operating Characteristic (ROC) area and Root Mean Squared Error (RMSE). From analysis and comparison with results from related literature, the Random Forest was found to perform best.